تعصب تغییر همگام و ارتباط فیزیولوژیک آن در بیماران مبتلا به اختلال پانیک

A covariation bias, i.e., the overestimation of random contingencies between fear-relevant stimuli and aversive consequences, seems to characterize anxiety disorders. Panic patients (n=30) and healthy controls (n=25) were exposed to panic-relevant, neutral, and phobia-relevant but panic-irrelevant picture stimuli, followed randomly be aversive consequences (acoustic startle stimuli). While covariation estimates reflected objective contingencies in both groups, only panic patients revealed a more negative Contingent Negative Variation (CNV) to panic-relevant than to phobia-relevant and neutral pictures. For startle reflex, only main effects of picture category were found, indicating that valence effects of picture stimuli were not specifically distorted in panic patients. CNV presumably reflects a biased processing of disorder-relevant stimuli by panic patients, perhaps with the expectation that aversive consequences will follow these stimuli.

Cognitive theories of panic disorder have underlined the role of cognitive processes in development and maintenance of the disorder (e.g., Clark, 1986). Panic patients were found to show an attentional bias (McNally et al., 1994) and lowered perceptual thresholds (Pauli et al., 1997) for fear-relevant word stimuli. Interpretational and memory biases are reflected in findings that panic patients compared to healthy controls, are prone to interpret ambiguous bodily sensations as threatening (Clark et al., 1988), and to remember anxiety-related situations (Becker, Rinck, & Margraf, 1994) and anxiety-related words (McNally, Foa, & Donnell, 1989; Pauli, Dengler, & Wiedemann, submitted) especially good. Finally, panic patients are assumed to be characterized by a covariation bias (CB), an overestimation of aversive consequences following panic-relevant stimuli (Pauli et al., 1996 and Pauli et al., 2001). In a typical covariation bias experiment, fear-relevant (FR) and fear-irrelevant (FI) picture stimuli followed by aversive or neutral consequences are presented to participants. Covariation estimates (CEs) for picture category–consequence combinations may be assessed before the experiment (pre-experimental or a priori CEs or expectancy estimates), during the experiment (on-line CEs or expectancy of consequences estimates), or after the experiment (post-experimental or a posteriori CEs). A covariation bias (CB) is reflected in enhanced CEs for the FR stimuli–aversive consequence combination compared to other combinations. While pre-experimental CBs have been found frequently in both anxiety patients and healthy controls (e.g., Amin & Lovibond, 1997; Kennedy, Rapee, & Mazurski, 1997; McNally & Heatherton, 1993), a post-experimental CB was rarely present in healthy controls, but was observed consistently in phobic patients (e.g., De Jong, Merckelbach, & Arntz, 1995; De Jong, Merckelbach, Bogels, & Kindt, 1998; Pauli, Wiedemann, & Montoya, 1998; Pury & Mineka, 1997; Tomarken, Sutton, & Mineka, 1995), in high fear subjects (e.g., Amin & Lovibond, 1997; Tomarken, Mineka, & Cook, 1989), and in panic-prone individuals (Pauli et al., 1996 and Pauli et al., 2001). It seems that low fear but not high fear individuals are able to correct a pre-existing CB on the basis of disconfirming situational information, and this is reflected in changing on-line CBs (Pauli et al., 1996). Similarly, the experience of relatively high contingencies between FR stimuli and negative consequences can induce a post-experimental CB even in low fear subjects (Pauli et al., 1996 and Tomarken et al., 1989), but not the experience of a high contingency between FI stimuli and aversive consequences (Pauli et al., 2001). However, an existing CB in anxiety disorder patients may be reduced by treatment (De Jong, Merckelbach, Arntz, & Nijman, 1992), and the efficacy of this change was found to be a significant predictor of long-term treatment success (De Jong, van den Hout, & Merckelbach, 1995). Covariation estimates are subjective measures, and therefore are highly susceptible to problems of experimental demand. Neurophysiologic measures of expectancy such as the Contingent Negative Variation (CNV) may circumvent this problem. CNV is a slow negative EEG potential shift presumably reflecting processes, such as expectancy, motivation, or response preparation (Birbaumer, Elbert, Canavan, & Rockstroh, 1990). CNV is typically elicited with a two-stimulus paradigm. After a first warning stimulus (S1), a negative moving slow brain potential can be observed reaching a maximum before the second, imperative stimulus (S2). For S1–S2 time intervals greater than 4 s, two CNV components can be distinguished: the initial or iCNV is modulated by the S1 information about the S1–S2 relationship; the terminal or tCNV is related to the motivational relevance and to the emotional quality of S2 (Rockstroh, Elbert, Canavan, Lutzenberger, & Birbaumer, 1989). Regan and Howard (1995) used CNV to examine expectancy of FR stimuli in healthy subjects. A warning stimulus (S1) announced presentation of FR stimuli (pictures of fear-inducing animals) in one experimental condition, and presentation of FI stimuli (pictures of landscapes) in another experimental condition. Two subcategories comprising half of the pictures were created within FR (two different types of animals, randomly chosen from five different categories, i.e., insects, snakes, dogs, cats, or rodents) and FI (lake or mountain scene) stimuli. Pictures of one subcategory within FR and FI stimuli served as conditioned stimuli (CS+), and were followed by an aversive startle tone, while the others (CS−) were not followed by any consequences. The CNV amplitude (determined in a time window 500–3500 ms after S1) in the FR condition was significantly greater for CS+ than for CS−. In the FI condition, no significant differences between CS+ and CS− surfaced. These results demonstrate that the stronger effect of conditioning for FR stimuli (compared to FI stimuli) results in enhanced CNV amplitudes for FR stimuli. Since distorted expectancies as reflected in CEs may also be associated with stronger conditioning (Davey, 1992), it might be hypothesized that anxiety patients show enhanced CNV amplitudes triggered by FR stimuli, indicating the biased expectancy of an aversive consequence following these stimuli. According to Tomarken et al. (1995), the CB of anxiety patients crucially depends on an affective match between FR stimulus and negative consequence. However, the large literature on affect-modulated startle response ( Lang, 1995) demonstrates that the emotional valence of picture stimuli and the emotional responses to aversive stimuli (e.g., startle stimuli) presented during or after picture presentation are not independent. First, as predicted by the motivational priming hypothesis ( Lang, 1995), the emotional valence of foreground picture stimuli modulates the responses to startle stimuli (e.g., Lang, Greenwald, Bradley, & Hamm, 1993; Vrana, Cuthbert, & Lang, 1989). Second, startle responses elicited during presentation of negative picture stimuli are higher in anxiety or panic patients than in healthy controls ( Cook, Davis, Hawk, Spence, & Gautier, 1992; Grillon, Ameli, Goddard, Woods, & Davis, 1994; Hamm, Cuthbert, Globisch, & Vaitl, 1997). Finally, the expectation of an aversive consequence may affect the response to it ( Pauli, Diedrich, & Müller, 2002; VanOyen Witvliet & Vrana, 2000). To assess these effects, it may be advantageous to use startle stimuli as aversive consequences in CB studies. Modulation of the startle response ( Lang, 1995) may objectively indicate the affective valence of the delivered stimuli depending on the presented picture. The present pilot study was designed to replicate and extend previous findings of a pre-experimental CB in panic patients (Wiedemann, Pauli, & Dengler, 2001) and a post-experimental CB in panic-prone individuals (Pauli et al., 1996 and Pauli et al., 2001). Disorder specificity was examined by using panic-relevant stimuli (medical emergency pictures) as compared to phobia-relevant but panic-irrelevant (spider pictures) and neutral stimuli (mushroom pictures). Furthermore, the CNV between picture onset (S1) and possible consequences delivered at picture offset (S2) was registered as an electrophysiological measure of expectancy. A more negative CNV was expected in panic patients but not in healthy controls in response to panic-relevant stimuli as compared to phobia-relevant and neutral stimuli. Aversive startle tones (white noise) were delivered at picture offset with a probability of 50% for all picture categories. Startle response amplitude served as a measure of the emotional valence, and panic patients were expected to show enhanced startle response amplitudes associated with panic-relevant pictures. On-line CEs were not assessed since motor responses to indicate expectancies during trials would have interfered with CNV recordings.